Sea Ice Type Classification in the Baltic Sea from TanDEM-X Imagery

Inhalt:
The current study uses the mean backscattering and coherence values of sea ice types for automatic sea ice classification using high resolution co-polarized TanDEM-X SAR images acquired the over coastal sea in the Gulf of Riga. Using the backscattering and coherence threshold values calculated for fast ice, thin smooth ice, pancake ice and water, an ice map of one SAR image was constructed. The best result for ice type discrimination was achieved when SAR images with a high incidence angle (>44deg) and monostatic imaging mode were used. The algorithm for ice type discrimination was effective at discriminating between selected ice types, although some areas (23.20%) were unclassified.